Voltage regulators, such as DC to DC converters, are used to provide stable voltage sources for electronic systems. Efficient DC to DC converters are particularly needed for low power devices. One type of DC-to-DC converters is a switching voltage regulator. A switching voltage regulator generates an output voltage by alternately coupling and decoupling an input DC voltage source with a load. The coupling and decoupling action can be performed by a switch, while a low pass filter comprising a capacitor and an inductor can be used to filter the output of the switch to provide a DC output voltage.
FIG. 1 shows an example implementation of a “buck” type switching regulator, which can perform a DC-DC down conversion. Referring to FIG. 1, circuit 100 includes a voltage source 103, a switching regulator 102 and a load 113. Switching regulator 102 is coupled to the voltage source 103 through an input terminal 114. Switching regulator 102 is also coupled to the load 113, which can be another electronic circuit that draws current, via an output terminal 112. Switching regulator 102 includes a switching circuit 116, which serves as a power switch for alternately coupling and decoupling input terminal 114 to an intermediate terminal LX node 109. Switching circuit 116 includes a first transistor 107 and a second transistor 108. Typically both transistors 107 and 108 can be implemented as metal oxide semiconductor field effect transistor (MOSFETs). Transistor 107 has a drain connected to input terminal 114, a source connected to an intermediate terminal 109, and a gate connected to a control line 105. Transistor 108 has a drain connected to intermediate terminal LX node 109, a source connected to a low voltage potential 115 (e.g. a ground), and a gate connected to a control line 106.
Switching regulator 102 includes a controller 104 to control the operation of switching circuit 116 via control lines 105 and 106. Switching regulator 102 also has an output filter 117, which includes an inductor 110 connected between intermediate terminal 109 and output terminal 112, and a capacitor 111 connected in parallel with load 113. Controller 104 causes switching circuit 116 to alternate between a first conduction period, where first transistor 107 is enabled and second transistor 108 is disabled to bring intermediate terminal 109 to a voltage substantially equal to the input voltage, and a second conduction period, where first transistor 107 is disabled and second transistor 108 is enabled to bring intermediate terminal 109 to a voltage substantially equal to that of low voltage potential 115. This results in a rectangular waveform, which toggles substantially between input voltage and a voltage equal to voltage potential 115, at LX node 109, which can act an intermediate terminal. LX node 109 is coupled to output terminal 112 via output filter 117. Output filter 117 converts the rectangular waveform at intermediate terminal 109 to a substantially DC voltage at output terminal 112. The magnitude of the output DC voltage at terminal 112 depends on the duty cycle of the rectangular waveform at intermediate terminal 109.
With widespread use of BCD (Bipolar-CMOS-DMOS) technology, it is common to integrate controller 104, switching circuit 116, as well as high precision feedback circuit (not shown in FIG. 1) on a single controller chip. The controller chip can have an output port that corresponds to the LX node 109. The controller chip can then be connected to a discrete inductor (e.g., inductor 110) at the output port that corresponds to the LX node 109 to form the switching regulator 102. The external inductor can also be connected to other discrete components (e.g., capacitor 111) to form an output terminal (e.g., output terminal 112) of the switching regulator 102.
FIG. 2 illustrates a way of placing the controller chip and the discrete components on a printed circuit board (PCB) to form the switching regulator 102 of FIG. 1. As shown in FIG. 2, system 200 includes a controller chip 202 which can include, for example, controller 104 and switching circuit 116 of FIG. 1. System 200 also includes capacitor 111, and inductor 110 of FIG. 1. FIG. 2 also illustrates a number of board traces and solder pads arranged to be LX node 109, vout 112, yin 114, and GND 115 of FIG. 1. System 200 also includes a capacitor 203 (not shown in FIG. 1) connected to yin 114 to act as a by-pass capacitor to further reduce the switching noise at that node. As shown in FIG. 2, controller chip 202, is placed adjacent to capacitors 111 and 203. Capacitors 111 and 203 are also placed adjacent to inductor 110.
The arrangement of the components in FIG. 2, while simple to implement, brings about a few drawbacks. First, such an arrangement takes up substantial board space, since each of the aforementioned components occupies a different area on a board surface. Second, relatively long board traces are needed to connect between the components, leading to huge parasitic capacitance at some of the critical nodes. For example, as shown in FIG. 2, the length of LX node 109 is about 3 cm. As LX node 109 an intermediate node for charging and discharging of inductor 110 and capacitor 111, by first and second transistors 107 and 108, reducing the length of LX node 109 and the associated parasitic capacitance can reduce switching loss and improve the efficiency of the power converter.
FIG. 3 illustrates an approach of component arrangement to form the switching regulator 102 of FIG. 1, for reducing the board space. As shown in FIG. 3, switching regulator 300 includes a controller chip 302 which can include, for example, controller 104 and switching circuit 116 of FIG. 1. Switching regulator 300 also includes inductor 110, which is housed inside an inductor housing 304. Inductor housing 304 also houses internal wires 309a-b, with wire 309a soldered to board 306 at solder pad 308a. Controller chip 302 also includes controller pads 307a and 307b which provide electrical connection to internal components of controller chip 302 (e.g., switching circuit 116 at LX node 109). To reduce the space occupied by switching regulator 300, controller chip 302 is disposed on top of inductor housing 304. Bond wire 310a is configured to provide electrical connection between controller pads 307a and solder pad 308a, thereby providing electrical connection between inductor 110 and controller chip 302 (e.g., LX node 109). Bond wire 310b is configured to provide electrical connection between controller pads 307b and solder pad 308b on the board. Solder pad 308b enables controller chip 302 to be electrically connected to other components on board 306.
While the arrangement shown in FIG. 3 reduces the board space required for placing controller chip 202 and inductor housing 304, the bond wires 310a and 310b are still relatively long, and can contribute to considerable amount of parasitic capacitance and resistance.
FIG. 4 illustrates an approach of component arrangement for addressing the drawbacks of FIG. 3. As shown in FIG. 4, switching regulator 400 includes a controller chip 402 which can include, for example, controller 104 and switching circuit 116 of FIG. 1. Controller chip 402 can be a flip-chip device and include solder balls 408a-b, which can act as terminals configured to provide electrical connections to, for example, controller 104 and switching circuit 116 disposed within controller chip 402. Switching regulator 400 also includes inductor 110, which is housed inside an inductor housing 404. Inductor housing 404 includes solder pads 410 configured to provide electrical connections to inductor 110 via internal wires 409a. As shown in FIG. 4, inductor 110 is electrically connected to controller chip 402 via solder ball 408a and solder pads 410. Inductor housing 404 is disposed on board 406. To reduce the space occupied by switching regulator 400, controller chip 402 is disposed on top of inductor housing 404.
With the arrangement of FIG. 4, where the electrical connections between controller chip 402 and inductor housing 404 are provided by solder balls and pads, the parasitic capacitance associated with those electrical connections (including those for LX node 109) can be reduced. However, there are still numerous drawbacks with the arrangement of FIG. 4. First, while the solder pads and solder balls provide good electrical connections between the inductor and the controller (e.g., the LX node), they are exposed to the environment, and noises can be coupled into the electrical connection. The noise can affect the voltage output of switching regulator 400. Second, with controller 402 separated from board 406 by inductor housing 404, the arrangement of FIG. 4 can degrade other electrical connections for controller 402, such as yin 114, GND 115, vout 112, etc. Long bond wires may still be needed to providethese electrical connections, which can contribute considerable amount of parasitic capacitance and resistance to those connections.
Therefore, while the arrangement of FIG. 4 improves the electrical connection between the controller and the inductor, it can degrade the rest of the electrical connections for the controller, and the performance of switching regulator 400 can still be compromised.
Hence, there is a need for a technique to arrange the components of switching regulator, not only to reduce the board area requirement but also to provide good electrical connections and good insulation for all of the components, such that the regulator can be made more compact and can be more easily fitted into devices of small form factors, such as mobile phones.